The point of requiring a second implementation is to prove that the standard is stable and consistent enough that at least two of them can agree on how it works

The point of requiring a second implementation is to prove that the standard is stable and consistent enough that at least two of them can agree on how it works. It doesn't matter who is building the second implementation, it just needs to exist in a feature complete and maintained state. If hobbyist implementations have been uncovering issues in the standard, that just further justifies the "two independent implementations" rule. Bug-for-bug compatibility problems are browser vendors' worst nightmare.